PICA-BASED ALGORITHM FOR AUTOMATIC DETECTION OF RESTING-STATE FUNCTIONAL NETWORKS. IMPLEMENTATION ON DIGITAL LAB PLATFORM

摘要

Understanding of the human brain architecture and its neuronal functional connectivity is an important neuroscience goal, because it may help to understand how the brain processes a large-scale information stream. Resting-state functional Magnetic Resonance Imaging (fMRI) is a popular neuroimaging tool that measures spontaneous low-frequency fluctuations in the BOLD (Blood Oxygenation Level Dependent) signal to investigate the functional architecture of the brain. In resting state one can reveal the co-activation of definite brain regions in distributed networks, called resting-state networks, which are selected by Independent Component Analysis (ICA) of the fMRI data. Although ICA decomposition in fMRI is widely used to identify networks, there is still no unique standard selection criterion to determine networks with potential functional connectivity. One of the main difficulties with component analysis is the automatic selection of the ICA features related to brain activity. In this paper, we describe an implementation of PICA-based algorithm for automatic selection of resting-state functional networks on Digital Lab Platform, including data processing on the Kurchatov Institute Supercomputer and Data Analysis Module, which can be used to detect neural networks and reduce subjectivity in ICA component assessment. In this work, rest-fMRI data sets were used, obtained on a Siemens Verio Magnetom 3T Tomograph of the Kurchatov Institute Resource Center.